[(*steve*)]

I've tried a couple of times.

It looks simple.

They're a 3 terminal regulator. (What could be easier)

But I find that they draw excessive current -- and with an input current limited to 10mA or so, the voltage is dragged down to under 2V.

It's almost like I've connected them around the wrong way.

I've tried the TO-92 and the SO-08 version (I was planning to breadboard with the TO-92 and implement with the SO-08) and I've seen the same thing with both.

I ended up using a 78L05 with its higher quiescent current.

 

[Harald Kapp]

We (not me personally) are using the LM2931 in the adjustable version without any known issues.
You have´the input and output caoacitors in place? Have you had a look at the output voltage using a scope? The regulator may be oscillating.
If the caps are in place, what type are they? I recently learned that some regulators expect an output filter cap with a certain minimum ESR (note: minimum, not maximum). If a capacitor with too low an ESR is applied, the regulators's feedback loop can become instable. Try tantalum or electrolytic capacitors, not ceramic (a small ceramic cap <= 100nF for HF rejection in parallel to a big tantalum capacitor is o.k.)

 

[(*steve*)]

Yes, input and output capacitors were in place.

I checked the ESR of the capacitors and they (the output caps) fell right in the middle of the stability range for the chip (they were electrolytic).

 

[Harald Kapp]

Mhhhm???

 

[(*steve*)]

Yeah. That's almost exactly what I thought.

At one stage I thought they didn't like a high impedance supply (e.g. when it was current limited) but although it *seemed* to work when connected to a battery, it later drew a large current from the battery too.

It's weird.

I've never had this much trouble from a three terminal regulator.

 

[TedA]

Steve,

The LM2931 regulators draw a lot of current during the rise of the input voltage, until the output reaches regulation. This is because the internal control circuit rather stupidly applies full base current to the PNP pass transistor, in an attempt to increase the output voltage.

So even though operation is fine once the regulator is regulating, it may never reach that happy state, if the input cannot provide the starting current.

See figure 23 of the current datasheet: http://www.ti.com/lit/ds/symlink/lm2931-n.pdf

The input current is shown as typically almost 25mA when the input is a small amount below the regulated output value. Even with no load on the output.

The part was originally intended to run off an automotive battery, not a low current source.

Some other manufacturer's LDOs do not have this "feature".

Due to the rather low gain of the PNP pass transistor, these regulators are not too efficient, even when they are regulating. The base current has to come from the common pin, rather than the output pin.

You might look for a CMOS part. Since the output device requires no drive currrent, they usually behave better while starting-up.

I was bitten by this one, years ago. DC power supply would not start-up. We did find another LDO regulator in the same package that worked OK with a pretty high DC source impedance. May have been by TI.

Ted

 

[(*steve*)]

*Thanks*

What a trap for young players!

It explains the problem exactly.

And the solution... Give it more curry!

 

[Harald Kapp]

Very interesting. One never stops lerning. Thanks.